Mining Meadows: Biosynthesis and Bioactivity of Phytochemicals from UK Plants

The process of domestication invariably leads to a loss of genetic diversity and a reduction in population size, leaving the animals more vulnerable to mutations. As these populations interbreed and mutations accumulate, they can cause disease, developmental disorders and infertility - making them ideal models for rare human disorders. Most studies focus on the ~2% of the genome encoding proteins, mainly due to the difficulty in identifying functional non-coding elements. This leaves aside the non-coding regions - sometimes called 'the dark genome' - which harbours functional regulatory sequences. The majority of disease- or trait-associated variants are present in these non-coding, often conserved, sequences.
The aim of the project is to use a unique resource consisting of 252 mammal genomes, along with population data, to assess the extent by which population sizes have been reduced and deleterious mutations accumulated. The latest developments in machine learning will be applied to transfer models and information from highly-studied organisms (human, mouse) to genomes with scarce resources, identifying functional elements and predicting the impact of mutations within them.